Accelerated Learning
— thinking out loud — 3 min read
The only way to win is to learn faster than anyone else. - The Lean Startup
Hi,
I am Aman, and this safe space is where I am most raw with my thoughts, hop on if you'd like to interact with them :))
Learning?
Learning is a very broad topic and I will try my best to be as narrow as possible, I intend this to be more of a prediction/thesis of what I think will happen in the future,
somewhat similar to
These are some of my personal bets about the future, It will be fun to come back to it after a few years.
Learning <-> Multiplayer
Learning has been multiplayer for quite sometime now,
In 2023 it's very hard to imagine learning as a single player game, we have done a very good job of exploring all possible models of [learning + community]
Asynchronous Multiplayer Experience
Collaboration exceeds the concept of timezones now and it's very hard to separate learning from community. For the most part web has positioned itself as an aggregator of communities and this model has worked out quite well.
- We now have "internet" friends distributed globally
- Opportunities beyond our geographical boundaries
- Accessibility to smart people
All this is amazing, but what excites me the most is the not so much about the community,
It for sure accelerates learning but my projection focuses more on the acceleration of our individual learning curves.
AI <-> Learning
Now the part that excites me the most: Imagining AI into our workflow,
I have briefly talked about some intersections of it in the past, here's a gist if you want to dive deep into the
"Think like Aman" rabbit hole:
Accelerated Learning
What tools like ChatGPT have enabled for us, is far beyond learning async with online communities, technology finally in some capacity has caught up to fiction.
A conversation based learning model similar to personal tutelage where I can get answers quickly without getting frustrated, sounds a bit vague but if thought about at depth, unlocks whole new sets of experiences.
Of course there is a nuance associated with any kind of information/knowledge and it's no different with ChatGPT.
The hard part is getting better at filtering out signal from noise,
getting better at things is hard and AI can only take a part of the load, there always has to be a human acting as a differentiator and It will be interesting to see technology acting as a bridge between good to great.
Asymmetry in Learning
People who are good at learning always have an edge whether it's sports or academia but now with AI in the loop, that bar has been reduced significantly.
A while back I wrote about the competitive edge that learning brings and it's interesting to see the formation of new models around this idea.
Yes sure, pessimists are worried about alignment and safety but the kid inside me who still has the curiosity to learn something new can't help but be more excited.
Proof of learning
I have talked about proof of curiosity and I think there soon will be a time where people will be able to showcase their proof of learning and project how good they are good at learning something.
Most platforms/companies geared towards solving hiring/education, revolve around solving for proof of work, but I personally think proof of curiosity and learning is a better model.
As the bar for learning goes down, learning something hard will not be a barrier anymore, how good people are at learning "x" will be more important.
When I think about companies who are in some sense trying to solve for such problems, Replit is what comes to my mind, excited to see more iterations of such a model.
No correct model
There is no correct model or idealised guide for how to learn x
Guides can take you up to a point, curiosity is the true deciding factor of how far you actually want to go, learning how to be curious is a skill in itself,
The barrier for learning and creating on the internet is going to have a significant reduction, where the concept of people belonging to different fields actively contributing to a field different from their domain will cease to exist as a question.
Personal Experience
Inspired by Toumas's talk about the linear sync engine, I tried to implement a dumb down version of it myself.
I have very little experience with GraphQl and mobx but ChatGPT did a pretty good job of giving me starting points.
Here's a little gist:
The sync engine implementation wouldn't have been possible without ChatGPT which makes me more excited for the future of work in general.
Conclusion
How to approach learning/building x is very clear to me now,
This blog will see more of my raw thoughts taking shape, iterations on the old one's and many more.
Definitely drop me a Hi, if I made you at least a little bit curious ;)
